From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jani Nikula Subject: Re: [PATCH] drm/i915: apply the PCI_D0/D3 hibernation workaround everywhere on pre GEN6 Date: Wed, 01 Jul 2015 15:42:51 +0300 Message-ID: <87oajwujzo.fsf@intel.com> References: <1435673207-23030-1-git-send-email-imre.deak@intel.com> <20150630172706.GA28262@amd> <20150701090207.GR5176@intel.com> <20150701095127.GB7969@amd> <20150701105331.GS5176@intel.com> <20150701123524.GE7969@amd> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <20150701123524.GE7969@amd> Sender: stable-owner@vger.kernel.org To: Pavel Machek , Ville =?utf-8?B?U3lyasOkbMOk?= Cc: Imre Deak , intel-gfx@lists.freedesktop.org, "Rafael J. Wysocki" , Ilya Tumaykin , Dirk Griesbach , Mikko Rapeli , Paul Bolle , Daniel Vetter , stable@vger.kernel.org List-Id: intel-gfx@lists.freedesktop.org On Wed, 01 Jul 2015, Pavel Machek wrote: > On Wed 2015-07-01 13:53:31, Ville Syrj=C3=A4l=C3=A4 wrote: >> On Wed, Jul 01, 2015 at 11:51:27AM +0200, Pavel Machek wrote: >> > > > > - Embedded panels have a well defined shutdown sequence. We >> > don't >> >=20 >> > > > > have >> > > > > any good reason to not follow this, in fact for some panel= s the >> > > > > subsequent reinitialization could be problematic in case o= f a hard >> > > > > power-off. (Thanks to Jani for this info) >> > > >=20 >> > > > Please cite concrete example. I have yet to see machine that w= ould not >> > > > power up on forced power down. In fact, I argue that such mach= ine >> > > > would be very broken, and that such machine does not exist. Wh= ile we >> > > > have these real machines broken: >> > > >=20 >> > > > > + * Lenovo Thinkpad X301, X61s, X60, T60, X41 >> > > > > + * Fujitsu FSC S7110 >> > > > > + * Acer Aspire 1830T >> > > >=20 >> > > > What makes you think that BIOS writers will do something diffe= rent for >> > > > Gen6+ hardware? X301 is not that old. >> > >=20 >> > > Thinkpad X1 Carbon (IVB) is perfectly happy with the D3, so clea= rly >> > > something changed for Lenovo at least. And most machines (old an= d new) >> > > have no problems whatsoever with the D3. >> >=20 >> > Well, one machine being happy does not matter that much... >>=20 >> I said most machines, not one. >>=20 >> > as going to >> > D3 has no real benefits. >>=20 >> Sure it does. Eventually we'll want to avoid resuming runtime suspen= ded >> devices when entering system suspend. For broken machines we'd need = to >> resume the GPU at that point. > > You want to optimize transition between suspend-to-RAM and > hibernation? No? I thought so. > > So no benefits, 7 real, broken machines. runtime suspended !=3D suspend-to-RAM > Pavel > > --=20 > (english) http://www.livejournal.com/~pavelmachek > (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/hors= es/blog.html --=20 Jani Nikula, Intel Open Source Technology Center